Geogr. Fis. Dinam. Quat. DOI 10.4461/GFDQ.2013.36.6 36 (2013), 81-94, 14 figg., 1 tab.

KENNETH HEWITT (*)

«THE GREAT LATERAL MORAINE», KARAKORAM HIMALAYA, INNER ASIA

ABSTRACT: HEWITT K., «The Great Lateral Moraine», Karakoram ured by non-glacial processes. A paraglacial influence is also present, Himalaya, Inner Asia. (IT ISSN 0391-9838, 2013). mainly through glacially induced rock slope instabilities. These lead to large postglacial landslides blocking rivers or descending onto smaller Large moraines and related ice margin deposits, observed along the surviving glaciers. The interpretation offered is a challenge for existing ablation zones of Karakoram glaciers, have been grouped together as the views of late Quaternary developments. Great Lateral Moraine (GLM). It was formerly attributed to the Little Ice Age. Other studies propose a longer sequence beginning with late KEY WORDS: Lateral moraines and troughs, Ice-marginal ramps, Mo- Pleistocene glaciations. All investigations have assumed the GLM records raine-dammed glaciers, Breach-lobes, Surge-type glaciers, Rockslide-rock climate-driven glacier expansions. Evidence presented here challenges avalanches, Fragmented drainage systems, Intermontane sedimentation, this, and the idea of a single origin in time or process. Bualtar Glacier in Transglacial processes, Valley glacier landsystems, Karakoram Himalaya. the Hunza Basin, with a much-discussed GLM, introduces the complexi- ties involved. The glacier is surge-type, its fluctuations affected by large landslides onto the ice. Both have triggered depositional episodes out-of- phase with surrounding glaciers and climate variability. More decisive has INTRODUCTION been local base-level control by landslides downstream of Bualtar, espe- cially the late Holocene, Baltit-Sumayar landslide. Similar conditions are Along the ablation zones of many Karakoram glaciers shown to affect many, if not all, GLMs. No consistent relations were bare cliffs in old lateral deposits rise from the ice edge and found with glacier size, morphology or known patterns of advance, but culminate above in prominent lateral moraines. The largest many surge-type glaciers and landslides in glacier basins are involved. A pervasive influence has been blocking of the upper Indus streams by of these were formerly regarded as a single regional phe- large mass movements. To address these complex developments, valley nomenon and called the Great Lateral Moraine (GLM). glacier landsystems concepts are employed, especially as applied to de- Meiners (1998, p. 55) describes it as «… a very well- bris-covered glaciers. Some distinctive Karakoram variants are identified. marked and well-formed lateral, partly high moraine, which The regional environment seems not to produce a unique type, but a surrounds the glacier tongues». A more explicit German complete spectrum of valley glacier landsystems. Recent evidence of glaciers transitioning between landsystem types suggest how GLMs have term is Ufermoranen-Dammen [«embankment moraine developed and why interactions of glacial, fluvial, lacustrine and eolian dam (or barrier)»] (Wiche, 1961; Haserodt, 1989, p. 212). systems, are important. GLMs are distinguished as «transglacial landsys- There are usually substantial troughs between the lateral tems», developments in which glacial activity is disturbed and reconfig- moraines and valley sides - the «ablation valleys» of older literature (Visser & Visser-Hooft, 1935-1938; Hewitt, 1993). In them, heterogeneous and discontinuous deposits (*) Department of Geography and Environmental Studies, Research build up where avalanches, rock falls and debris flows Associate, Cold Regions Research Centre, Wilfrid Laurier University, Waterloo, Ontario, N2L 3Z9, Canada. E-mail: [email protected] come from the valley slopes, and where drainage is chan- The paper is dedicated to Prof. Monique Fort, to acknowledge her spe- neled or impounded (fig. 1). cial and outstanding contributions to the geomorphology of the High Asian In the 19th and early 20th centuries, glacier ice was mountains. She has been an inspiration in tackling problems of scale and commonly observed standing at or above the lateral complexity that distinguish the region; that require attention to its peculiar- ities, and thinking «outside the box» of predominantly Eurocentric ideas moraines, adding to them and shedding water and debris (Fort, 1987, 1995; Fort & Peulvast, 1995; Fort, 2000). Investigations at into valley side troughs. It is something rarely observed Bualtar Glacier were funded by the International Development Research since the 1920s except during glacier surges, making it Centre, Ottawa, Canada; Pakistan’s Water and Power Development Au- seem logical to identify the GLM with the Little Ice thority; and Wilfrid Laurier University’s Office of Research. Local residents Age (LIA). Some equated it specifically with the «1850 and mountain guides, especially Mr. Shaffi Ahmed of Nagar, gave invalu- able information and field assistance. I am indebted to two reviewers for moraines» in the European Alps. According to von Wiss- helpful comments, and Ms. P. Schaus for preparing the figures. mann (1959), «… in High Asia [generally] the moraines…

81 FIG. 1 - The GLM along the left flank, mid-ablation zone of Chogo Lungma Glacier. The marginal trough is to the left of the great lateral moraine. Active ice to the right is at a somewhat higher level, but below the GLM crest (photo: K.H., 2003).

originate in the high stand glaciation at the middle of the lower tongue and for some 10 km up-valley are huge later- past [19th] century…». However, Kick (1989) challenged al moraines (fig. 2). Substantial slope, kame terrace, and this view of the «large moraine» at Chogo Lungma and other Karakoram glaciers and supported Mason’s (1930) view, that «… the majority of glaciers in the region were in a condition of maximum advance between about 1905 and 1915». Haserodt (1984, p. 83) argues, on the basis of tree ring data, for: «… a minimum age of the GLM of Bagrot [Glacier, near Gilgit] of 280 years… closer to the beginning of the “Little Ice Age” period…». However, since the mid- 1980s, most studies invoke several glacial expansions to explain these features, including some much earlier than the LIA (Schneider, 1969; Haserodt, 1989; Kick, 1989). While not using the term, Kalvoda (1992) attributes lateral margin deposits called GLM by others to events beginning in the late Pleistocene. All existing interpretations at least agree that the GLM records climate-driven glacier fluctuations. Evidence as- sembled here supports a diversity of origins, in time and process. Perhaps the term GLM should be abandoned. However, it identifies widespread, conspicuous ice-margin features in the region, is an important notion in the litera- ture and, in itself, does not imply a particular explanation. Here, the plural is used to reflect the variety of GLMs. Quotes are applied when citing other usage. An example introduces the conditions of interest.

THE BUALTAR GLMS

Bualtar Glacier in the Hunza Basin, sometimes called «Hopar Glacier», has an area of 115 km2 and a main ice stream 22 km long. It is largely avalanche-fed and drains northwards from a precipitous source zone below Diran FIG. 2 - Ablation zone of Bualtar Glacier looking downstream showing its GLMs (arrows), the associated valley side deposits, and steep cliffs to Peak (7,266 m) in the Rakaposhi Range. With a terminus the present ice edge. Ultar Peak (7,388 m) is in the top right background at 2,450 m, total basin relief is 4850 m. Surrounding the (photo: K.H., 1986).

82 lacustrine deposits fill valley side troughs. Similar GLM ing from the distinct advance of this glacier…». He adopts a features occur along the Barpu Glacier, which terminates much-expanded time frame, proposing that, «The highest beside Bualtar and was joined to it in the past (fig. 3). [moraines]… formed during the Hunza phase of glaciation in the upper Pleistocene». In a more recent paper, Kalvoda Existing studies and interpretations & Goudie (2007, pp. 112-115) assign the «huge walls of lateral moraines [close to Nagar village]…» to «… the last Bualtar and Barpu Glaciers appear in several influential advance of the valley glacier in the upper Pleistocene investigations of the «GLM» and related matters. Haserodt [62,000-68,000 years ago]» (p. 108 and 113). There are no (1984, p. 93) discussed the GLM at «Hoppar» and sketched actual age-determinations for GLM sediments or surfaces its cross-profile, placing the highest ice in the «17th, 18th or at Bualtar. Tree ages noted by Haserodt (1989), as earlier 19th c.». A later paper describes the features as «High Stand by Wiche (1958), were not supported by tree-ring or other moraines» of the LIA, defined by «300, 75-100, and 10-25 dating methods (Kick, 1989). Chronologies are based on year vegetation growth» (Haserodt, 1989, pp. 214-217). presumed morpho-stratigraphic relations and inferred ele- The same features are identified by Kalvoda (1992, vation and vertical relations to glaciations of the Hunza Plate XXVI/2, p. 189) as: «… Glacigenous sediments of the valley (Shroder & alii, 1993, p. 154; Owen, 2006, p. 15). lower part of the Bualtar valley-glacier tongue… Huge, in The state of the Bualtar GLMs in recent decades is af- some places up to 160 m high walls of lateral moraines dat- fected by erosion, or burial by wind-borne dust, and tram-

FIG. 3 - Topographical map of Bualtar and Barpu Glacier basins showing the extent of the GLM complex and debris-covered ice.

83 pling by animals and humans. More important, the great cliffs in former glacial deposits are «erosion» features. They do not record actual ice contact surfaces or trim- lines of the highest, last, or any past glacier expansion, but an extended period of degrading of GLM deposits (fig. 4). In particular, large slump blocks continue to form and slide towards the glacier on the Hopar and Shishkin sides (MacDonald, 1989; Hewitt, 2009a). Erosion has removed over half the original sediment build-ups above existing ice levels. The original GLMs and Bualtar ice surfaces must have been tens of meters higher than today’s rem- nants, and closer to the valley center (fig. 5). Conditions observed in recent glacier advances also need emphasis.

Bualtar fluctuations Reports describe intermittent sudden advances of Bual- tar, alternating with large retreats, sometimes stagnation FIG. 5 - Present-day transverse cross section of Bualtar Glacier, 5 km (Conway, 1894; Workman, 1908). Rapid advances were re- above the terminus to illustrate relations of present ice to GLM rem- nants. Depth of ice here (maximum 230 m) is based on monopulse radar depth soundings. (Vertical exaggeration, x 3).

ported in 1922 and 1923, and in 1929-1930 (Visser, 1928; Mason, 1930). Two separate, rapid accelerations occurred in 1987 and 1990 (Gardner & Hewitt, 1989; Hewitt, 2009a). They lasted a few months, causing massive distur- bance of the glacier and confirmed that Bualtar is a surge- type glacier (Hewitt, 1969). In both recent events, fast flow stalled before reaching the terminus, but the glacier advanced gradually thereafter, and further than at any time since the 1920s (Mason, 1930, 229-301). The surges triggered or accelerated large-scale mass movements in GLM materials (fig. 6). In surge-type glaciers, thickening and advances arise from surges in a cycle unique to each glacier (Sharp, 1988; Jiskoot, 2011). They complicate mass balance and relations to climate. In the late LIA, when most nearby glaciers were advancing, Bualtar retreated. Recently, when most others were retreating, it advanced. It is unlikely Bualtar only became surge-type in recent centuries when rhythms show roughly two episodes and four surges per century. If at all representative, these indi- cate 200 and 400 Holocene events, respectively. Surge ac- tivity could help explain the massive moraine building and other GLM features discussed below. Conditions are further compounded by landslides. In 1986, massive rock slope failures deposited about 20 Mm3 of debris onto Bualtar (Hewitt, 1988). By 2012, the debris sheet, some 4.5 km2 area, had been transported 10 km down-glacier, while suppressing ablation and standing 10- 20 m above the surrounding ice. Rock avalanche material is shed from the raised glacier surface. It forms ridges of lateral moraine that rise, fall or pinch out irregularly along the ice-edge (fig. 7). The deposits are «moraines» in geom- FIG. 4 - View from the GLM crest down the left cliff face to the lower etry and depositional style, but their composition and tim- Bualtar. Large rotational landslips carry former GLM deposits to the ing depend upon rock avalanche materials. The landslides, glacier margin. Clearly, the blocks in the foreground were much higher than the existing, eroded crest. Note the abandoned old road, paths and lasting a few minutes, created a disturbance continuing irrigation channels on top of the nearest landslips (photo: K.H., 2006). for decades and producing distinctive deposits (Hewitt,

84 FIG. 6 - Bualtar Glacier during the 1987 surge. The flood of ice and severe crevass- ing can be compared with 1986 in fig. 2. There was increased landslide activity along the margins and numerous outburst floods from ice margin lakes (arrow), all helping erode GLM materials (photo: K.H., 1987).

FIG. 7 - Bualtar ice margin during passage of thickened ice showing (high stand-type?) moraine-building fed by rock avalanche debris (photo: K.H., 2010).

2009a). The relevance to GLMs, of course, depends on alii, 2011a). Even so, neither the landslides, surges, or the landslide frequency. neoglaciation formerly invoked, offer sufficient explana- Local accounts describe a similar landslide onto Bual- tion of the GLMs. tar Glacier in the late 19th century and are supported by field inspection of the source area. In the 1980s, sand avalanches from the GLM cliff at Hopar had sedimentary GEOMORPHIC SETTING OF THE BUALTAR GLMS characteristics of rock avalanche matrix materials, appar- ently from a prehistoric event (fig. 8). In Barpu basin rem- Bualtar and the Hispar valley nants of three prehistoric rock avalanches have been found that descended onto the glacier (ibid.). The largest, from As Kalvoda & Goudie (2007) observe, the huge ter- near the summit of Spantik Peak (7,027 m), travelled 11 races of sediment on the Hispar River left bank, below the km down-glacier. Debris from it caps GLM ridges for an- Bualtar junction, seem related to its moraines. So are land- other 5 km at least, apparently emplaced in a surge event forms around the junction of Bualtar and the Hispar, (Hewitt, 2002, p. 368). Thus, landslide forcing seems a re- where multiple shifts in the glacier and stream channels curring source of moraine-building episodes (Hewitt & are recorded. Several superimposed or «epigenetic» rock

85 FIG. 8 - Sand avalanches from the flank of the Bualtar GLM below Hopar villages. Dozens were observed in 1985 and 1986. The composition, combining a single lithol- ogy and angular or very angular clasts, sug- gests prehistoric rock avalanche material deposited in that horizon of the GLM (photo: K.H., 1986).

gorges occur, where streams were let down onto former 3.5 km. Lateral moraines spread outwards tens of meters valley flanks or bedrock spurs from valley fill, landslide de- as well as building 5-15 m vertically within the old GLMs bris, or moraines (Hewitt, 1998; Ouimet & alii, 2008). (Hewitt, 2009a). Another discovery seems key to the scale Presently, Hispar River enters the Bualtar Glacier fore of locally confined GLM sedimentation. field and together with its outlet stream, rejoins a former Hispar epigenetic rock gorge, choked above by talus. The Baltit-Sumayar landslide Bualtar ice has dammed the Hispar, and outburst floods are reported from the glacial lakes (Workman, 1908). A large, late-Holocene landslide blocked the Hispar Landslides have dammed the river up-valley of the Bualtar River at its junction with the Hunza. A massive rock slope entrance. failure descended from above Baltit into and across the A more surprising feature is how closely the GLM Hunza River, also damming the Hispar and Silkiang rivers location and morphology conforms to today’s ice tongue (Hewitt, 2001). Deposits record lakes impounded for (fig. 9). Ice surface levels up to 250 m higher are not decades if not centuries. The terrace sediments on which matched by evidence of comparable advances or lateral the Nagar villages lie were likely deposited behind or pro- expansion of the ice (fig. 10). In recent years, a mere 10-15 tected by the landslide in the older glacial valley of the m thickening of the lower Bualtar involved an advance of combined Barpu-Bualtar, Gharesa and Hispar glaciers.

FIG. 9 - Satellite image of Bualtar glacier tongue showing close relations of GLMs to existing Hispar and Barpu ice streams (source: landsat ETM+ image, acquired 2006-07-26, covering a SRTM DEM, reso- lution 90 m). Location of the cross-section in fig. 5 is shown. Remains of the 1986 rock avalanche debris are at and below right hand arrows. Vegetated areas are irrigated land of Hopar villages (left hand arrow).

86 FIG. 10 - View looking north east from above Hispar River, showing the vertical extent of the Bualtar GLM deposits and how they wrap around the existing, lower Bualtar tongue (arrows). The treed valley side trough at the right of the photograph was occupied by the breach lobe in fig. 13 (photo: K.H., 2001).

The Baltit-Sumayar deposit is 8-10 km downstream of the Relations to Baltit-Sumayar are complicated by other Bualtar junction, but the dam crest is at a similar height to landslides that have blocked Hunza valley. There are at the highest GLMs (fig. 11). These morphological relations least fifteen large, post-glacial rock slope failures be- suggest an alternative explanation of sediment build-up tween its junction with the Gilgit and Sost (Hewitt, around Bualtar terminus. In its present position the Bual- 2001). Ghulkin and Gulmit glacier tongues lie at the tar tongue seems likely to have floated in the landslide lake head of a river reach that has been repeatedly blocked until sediment build-up filled the lake and elevated the ice. and aggraded, as in the 2010 Atabad landslide. Two pre- The Baltit-Sumayar event affected developments along the historic events in the adjacent Sarez section involved lower Hispar for several millennia. High stand conditions long-lived lakes. The uppermost lake beds over the No- lasting some centuries would give ample time for glacier mal megaslide close to Gilgit are at elevations of 2,500- surges and landslides to influence moraine building. Sub- 2,600 m, meaning its lake could have reached the Bual- sequently the landslide dam was breached, but gradually, tar terminus and Batura’s. There is no age-determina- and is not fully cut through. It remains local base level for tion, but the landslide post-dates the last major glacia- the Hunza and Hispar valleys. tion (Hewitt, 2001, 2009b).

FIG. 11 - Schematic longitudinal cross-section of the lower Hispar valley from Baltit-Sumayar land- slide dam to the Bualtar and Barpu GLMs.

87 INTERPRETATIONS This is a comprehensive approach, but requires qual- ification in the Karakoram context. Debris-covered ice The origins of the GLMs involve two interrelated sets and huge lateral margin complexes are widely present, of causal factors: ice-margin and valley side depositional but moraine-dammed termini like the example they processes, and controls capable of altering ice levels, dy- give of Ghulkin Glacier, much less so. Many heavily namics and, perhaps, debris supply. The first set of causal «covered» Karakoram glaciers in fact belong to Benn & relationships can be addressed in terms of the «landsys- alii’s (2003) «coupled», «outwash-head» types, includ- tems» of Benn & alii (2003). The second, involving exter- ing Bualtar now and including most of the largest ice nal controls, require some modified and extended «con- masses like Siachen and Baltoro. Their snouts overlook ceptual relationships», to use their terms. powerful outwash streams and there are few or no frontal moraines. Up-valley some, like Bualtar, Panmah and Debris-covered valley glaciers and landsystems Chogo Lungma, have well-developed GLMs. In others, like Baltoro and Sarpo Lago, they are poorly developed Landsystems, as they apply to valley glaciers, have been or absent. developed by Boulton & Eyles (1979), Eyles (1983), and Conversely, steep tributary glaciers with clean ice, or Benn & Evans (1998, Part 2). Appropriately, for GLM only irregular patches of debris, can be «moraine-dammed» concerns, the focus has been on glacial sediments and de- (fig. 12). Whether, and to what extent, debris builds up on posit assemblages. Those at glacier margins depend, firstly, the glacier surface depends as much on steepness, icefalls on the ratios of sediment and ice reaching them, and the and crevassing, as the ratio of ice and debris supply. efficiency or otherwise of transport through and beyond If largely neglected, there are thousands of rock gla- pro-glacial areas. In these terms, Benn & alii (2003) pro- ciers in the Karakoram, many of them glacier-derived pose a spectrum of landsystems, from clean or «uncovered (Owen & England, 1998; Shroder & alii, 2000). They glaciers», through increasingly heavy mantles in «covered» differ from the moraine-dammed glacier type in having glaciers, to rock glaciers at the other extreme (ibid., p. 375, tip-heap margins, usually continuous with surface ridge- their figs. 15.23). and-trough features, not a sharp transition between active Uncovered glaciers promote «… coupled ice mar- ice and moraines (Whalley & Martin, 1992; Humlum & gins…» with efficient transfer of sediment between the alii, 2007). There is exceptional variety of sizes and genesis glacier and pro-glacial fluvial systems. Lateral and termi- (Giardino & alii, 2011). Much larger examples occur than nal moraines tend to be minimal. By contrast, debris- are reported from intensively studied mountains elsewhere covered glaciers are portrayed as «decoupled», with high (Haeberli, 1985; Barsch, 1988, 1992). sediment supply, inefficient or absent removal at the All the terminus types proposed by Benn & alii (2003) margins, particularly relevant for the «giant build-ups» are found in the Karakoram. A unique landsystem is not associated with GLMs. Heavily debris-covered, «moraine- associated with its climatic regime or high mountain set- dammed» glaciers, are singled out as typically Hima- ting. Other conditions intervene to create a diversity of layan. In general, «covered» glaciers are identified with types, including the glacier surges and landslides men- high mountains and/or drier climates. Changes in landsys- tioned above. Furthermore, it is evident that changes in tems are said to be responses to climate, especially in- landsystem type need not depend upon climate change or creasing aridity or humidity. glaciations.

FIG. 12 - A tributary glacier in Nangmah valley, Hushe Karakoram; an «uncovered glacier» that is «uncoupled» and has a «moraine-dammed» terminal complex (pho- to: K.H., 2012).

88 Transitional types large landslide onto the glacier (Hewitt, 2001), carried lobes of terminal ice through the bounding moraines and Bualtar’s lower tongue has been debris-covered since down their steep forefronts. Most of the neighbouring the earliest modern observations, but is outwash-head or glaciers in Hunza have giant GLMs but are also outwash «coupled» type. By contrast, the GLMs record former head types. They involve another distinctive characteristic «giant bounding moraines» and «moraine-dammed» mar- of the Ghulkin-type; ice levels elevated by sub-glacial de- gins. Typical build-ups include the «repeated superposi- position, not necessarily dependent on ice thickness or tion of moraines around the margins» as described by confinement behind moraine-dammed margins. Benn & alii (2003, p. 384). The GLM deposits record typ- Gulmit and Ghulkin also sit up on ramps of sediment ical combinations of the «dump moraines and ice-margin behind moraine dams, at the head of an aggraded river aprons», the «ramps and fans» described by Benn and Evans (1998, pp. 475-480). Kuhle (1990) identified other reach. Bualtar, however, has eroded its bed to a level relevant processes, as did Shroder & alii (2000) on Nanga where the outlet stream crosses bedrock to join a degrad- Parbat glaciers. In each case, debris-mantled ice is consid- ing main valley that is deeply entrenched in valley fill. The ered important. It suggests that, when the high stand right flank below the active terminus is complicated by the GLMs were being built, the moraine-dammed type and re- large mass of stagnant ice related to the recent surges. lated features were more common. Kuhle’s (1990) discussion of «ice-marginal ramps» The «Ghulkin-type» of Owen (1994), heavily mantled deals with moraine-dammed, «decoupled» tongues where and moraine-dammed, is proposed by Benn & alii (2003) debris accumulates underneath as well as on and around as typical of debris-covered glaciers. Karakoram examples ice margins, similar to Ghulkin Glacier. «Breach lobes» include the Gulmit and Barpu, and the Lupghar and Kun- can be a singular indication of such conditions, where ice ti Glaciers whose GLMs are described by Meiners (1998). bursts through lower or weaker bounding moraines to This type seems important for GLM interpretation. How- emplace secondary lobes (Benn & alii, 2003, p. 384-385; ever, on the one hand, as noted above, such conditions are Deline, 2009). Other examples are recorded in moraine actually quite rare today. On the other hand, recent be- formations near the Ghulkin and Ghulmit termini, and haviour of Chillinji Glacier on the upper Karambar, or hundreds can be found in satellite coverage of Karakoram Yazghil in Shimshal, suggest this landsystem type is, at glaciers. Moreover, there is evidence of some five breach least partly, a response to other and special conditions, not lobes along the left flank of the Bualtar GLM - intriguing- only the debris cover. They may be fairly unstable. These ly, opposite where Barpu could have joined it. The highest two were Ghulkin-types in the recent past, Chillinji until of these also marks the largest distributary path of Bualtar, 2002. Both have shifted to coupled, outwash-head types, possibly involving both Barpu and Bualtar tongues (fig 13). mainly through «breach lobes» that overwhelmed moraine It is 150 m above today’s ice, at the same level as today’s dams (see below). The transitions observed seem relevant lower Barpu. It confirms that the original GLMs were for interpreting Bualtar GLMs. Relatively small increases much higher than today’s remnants, and the ice stream(s) in ice thickness and velocity, in Chillinji’s case following a elevated on an aggraded bed (fig. 14).

FIG. 13 - Moraines and channel of a breach lobe on Bualtar GLM left flank, 3 km from the present-day terminus. This and other examples suggest the ice lay in «ice-mar- ginal ramps» (Kuhle, 1990), was «moraine- dammed» (Benn & alii, 2005), and at a higher level than existing GLM remnants. Note the large landslips in GLM materials, collapsing towards the present ice surface (photo: K.H, 1987).

89 beyond today’s, Karakoram glaciation appears radically different from most of High Asia as well (Porter, 1970; Fort, 1995). The evidence has been challenged (Kuhle, 2004, 2006). The Baltit-Sumayar deposit enters this picture as «moraine». Schneider (1959, p. 207) calls it a «hardened rubble» formation (= «verfestiger Blockschutt»). It is a «deformed till» in Derbyshire & alii (1984, p. 488), and «morainic conglomerates» in Kalvoda & Goudie (2007, p. 109). Deposits upstream are attributed to glacial lakes of the «last main [Pleistocene] advance» (Goudie 1984, p. 383). The events are placed in the «Bhorit Jheel expan- sion» between 50-65 ka (Li & alii (1984), or 54.7-43.2 ka (Derbyshire & Owen, 1990; Owen & alii, 2002). It inter- prets the deposits as quite separate from, and much older than, the Bualtar GLMs. These dates imply an extraordinary slowing or cessation of erosion. The Baltit-Sumayar deposits were emplaced over fluvial gravels and much deeper, buried valley fill, still intact. It implies that, in 40+ ka, «no net» erosion has oc- curred; not in pre-existing valley fill let alone in bedrock. The Hispar between Bualtar and the Hunza flows over val- ley fill pre-dating the «Ghulkin I stade» which precludes net erosion there for 20+ ka (Kalvoda & Goudie 2007, p. 112). It means all the sedimentation and trenching that shapes the GLMs and valleys below them, has occurred above valley floor levels from the late Pleistocene, and well above the last phase of valley incision. It should be recalled that the region has globally FIG. 14 - Schematic representation of sequences involving neoglaciations versus a «transglacial» sedimentation/trenching hypothesis. The former exceptional relief and some of the highest known rates of requires massive and unlikely thickening of the ice without much lateral uplift and denudation (Searle, 1991; Park & alii, 2001; expansion or ice advance. The latter involves changing sub-glacial as well Zeitler & alii, 2001). The Hunza valley here has been as lateral sedimentation, and requires no great changes in ice thickness. called «the steepest place on Earth» (Miller, 1984). Other The reality may well involve some combination of both types of control, but with ice thickening and advance during surges. evidence suggests exceptional rates of denudation and pre- sent-day geomorphic activity (Shroder & alii, 1993; Bur- bank & alii, 1996). However, the glaciations view means that for 50,000 years or more, and while hundreds of me- It is suggested that the Bualtar GLMs record episodic ters of late Quaternary uplift occurred, there was no or transitional behaviour between aggraded, «Ghulkin- equivalent incision. type» and degraded outwash-head type landsystems. This The same view also invokes «much more» ice to ex- does return us to the broader question of how such plain the huge GLM deposition. At Bualtar and elsewhere, changes could come about. Equally intriguing is how both trenching and removal of that debris are identified with types occur in «covered» glaciers, in close proximity and, major glacier thinning and present-day conditions. It today, at the same time. As indicated, hitherto GLMs have seems to run counter to what is usually thought to been attributed to glaciations, and reasons for challenging strengthen or weaken glacial processes. this must first be addressed. The re-interpretation of the Baltit-Sumayar deposit as a landslide, not «moraine», is based on diagnostics from The glaciation hypothesis lithology and morphology (Hewitt, 1999). Granitics and granodiorites from the Ultar Massif above Baltit, comprise The main, «high stand» Bualtar GLMs have been 100% of the main deposit, and are emplaced over Hunza placed in the «Ghulkin I stade», described as an «ex- schists and carbonates of the Karakoram Metamorphic panded foot», or «minor valley glaciation» with an age Complex. The latter outcrop along both flanks of the of a 25.7-21.8 ka (Kalvoda, 1992; Shroder & alii, 1993, Hunza and Hispar valleys (Searle, 1991, p. 111). A trunk p. 154; Owen, 2006, p. 15). It coincides with the last ma- glacier descending either valley would carry and deposit jor glaciations of the Northern Hemisphere variously some, if not mainly, metamorphic debris. Other rock named «Wisconsinian/Würm/Weichsel» and when the avalanche criteria include distinctive broken and crushed Laurentide Ice Sheet reached almost to New York, and clasts in all size fractions, long run-out and run-up mor- most of Scandinavia and Britain lay under ice. If it is true phologies, while characteristics definitive of moraine are that ice limits in the Hunza Basin were only minimally absent (Hewitt, 2001).

90 Recently, the landslide event has been bracketed by a TABLE 1- Forty Karakoram glaciers with GLMs identifying those with well-constrained 10Be terrestrial cosmogenic nuclide age. developments similar to Bualtar Glacier. This is a very small, preliminary sample but sufficient to point to a diverse group across the Karakoram Quartz-rich rock, retrieved from boulders on the landslide surface has an exposure age of 4.36±0.14 ka (Hewitt & RAs Glacier GLACIER BASIN Surges1 RA dam GLOFs alii, 2011b). It implies zero net erosion for barely one- on ice block tenth of the glacial time frame, and massive aggradation N. Terong Nubra ?XXXX and trenching associated with the landslide, more in accord Kondus Saltoro XX with regional geomorphic activity. However, the signifi- Charakusa Hushe X (T) XXX cance of these findings also depends on how far landsys- Aling Hushe X (T) X X X X tem features at Bualtar apply elsewhere in the region. Baltoro Braldu X (T) X X Panmah Braldu X (T) XX Biafo Braldu X (T) X X Kutiah Stak X XXX OTHER KARAKORAM GLACIERS Mani Phu’gam ? X X X Hinarche Bagrot ? X X Virjerab Shimshal ? XX GLM constraints and correlates Khurdopin Shimshal X XXX Yazghil Shimshal ? ? X X X Barpu and its GLMs have been affected by the same Malangutti Shimshal XXXX constraints, if partially buffered by Bualtar. Tributaries of Koz Yaz Chapursan ? XX Barpu’s Sumaiyar Bar branch are surge-type, possibly the Yashkuk Y. Chapursan ? XXX whole glacier. The large mass of dead ice and thermo-karst Kuk-i-J. Chapursan ? XXX forming its moraine-dammed terminal lobe may derive Murkhun Hunza ?XX X Batura Hunza ? (T) XXX from a late 19th century surge (Conway, 1894). Several Pasu Hunza XXX large landslides onto Barpu were identified earlier. How- Ghulkin Hunza XXXX ever, this is a very local comparison. Hispar Hunza X (T) XXX Bualtar shares a number of relevant landsystem con- Garumbar Hunza X XXX Barpu Hunza X X X X X ditions with, for example, the Siachen, Kondus, Baltoro, Bualtar Hunza X X X X X Chogo Lungma, Hispar, and Bagrot glaciers. In addition Silkiang Hunza ? X to well-developed GLMs, their termini are: (i) Debris- Karambar Hunza X XXX covered: heavy supraglacial debris occurs over the lower Minapin Hunza X 15 km or more of each glacier tongue, but none is «de- Kukuar Hunza X XX Jaglot Hunza X XX coupled»; (ii) Outwash head type: all examples have this Shani Naltar X style of terminus with powerful pro-glacial streams and Kutu N. Naltar X no or limited terminal moraines; (iii) GLMs were em- Bhurt Karambar X placed over valley fill: each terminus has advanced and Karambar Karambar X X X retreated over valley fill tens to hundreds of meters thick. Pehkin Karambar XX Chillinji Karambar X X X X X Ice levels, terminus behaviour and pro-glacial sediment Chatteboi Karambar ? X X X removal have been constrained by downstream aggrada- Karambar P. Karambar XXXX tion or trenching, as reflected in; (iv) Relations of stream Chiantar Yarkhun X terrace and GLM geometries: downstream of the termini Pechus Yarkhun ? XXX stream terraces have upper levels generally lower than 1 «X» identifies an evident influence on GLMs; «T» refers surge of tributary of glacier the high stand GLMs, but continuous with associated named; «?» suggests possibility based on indirect evidence; and blank no evidence for glacier margin troughs; (v) Landslide interrupted drainage: an influence. in every case, one or more cross-valley landslide barriers exist a few kilometers downstream of the termini (He- witt, 1998, 2006). leys downstream have been blocked by one or more cross- Other well-known examples such as Biafo, Batura, valley landslide barriers (Hewitt, 1998, 2006). Yashkuk Yaz and Pechus may seem different. Their GLMs include extensive terminal moraines marking for- «Transglacial» Landsystems? mer and present ice positions. However, each has a low angle ice tongue entering a main river valley. Terminal GLM features of the Bualtar occur widely in larger moraines sit on top of river terraces that continue down-, Karakoram valley glaciers; most importantly a particular and up-valley of the glacier tongue. In these cases too, form of «coupling» of sedimentation between glacial and glacial deposition matches aggradation levels in pro-glacial fluvial systems (Hewitt & alii, 2011a). The GLMs are com- streams, as in item (iv) above. The other four conditions posed largely and uniquely of glacigenic sediments. High also apply to these glaciers. stand moraines and other final touches may record cli- An inventory of forty glaciers across the Karakoram mate-driven expansions, or surges, or pulses of debris fol- and with well-developed GLMs extends the perspective lowing landslides onto the ice. However, in all examples (tab. 1). A third involve surge activity, and massive rock investigated the timing, heights and volumes of GLM fea- slope failures are known in a quarter. In all cases, the val- tures are not simply outer or high stand markers of glacier

91 advances. They cap massive sediment build-ups extending REFERENCES and controlled beyond the ice fronts. This distinguishes them from conventional latero-terminal deposits. It ex- BARSCH D. (1988) - Rock glaciers. In: Clark M.J. (Ed.), «Advances in Pe- tends the range of valley glacier landsystems to include re- riglacial Geomorphology». Wiley, Chichester, 69-90. sponses to non-glacial conditions. BARSCH D. (1992) - Studies and measurements on rock glaciers at Macun, When other earth surface processes significantly influ- Lower Engadine. In: Evans D.J.A. (Ed.), «Cold Climate Landforms». ence glacial processes they generate what may be termed Wiley, Chichester, 457-473. «transglacial» landforms or sediment assemblages. «Trans- BENN D.I. & EVANS D.J.A. (1998) - Glaciers and glaciation. Arnold, Lon- glacial» has been applied especially mass movements along don, 734 pp. glacier flanks (Iturrizaga, 2006, 2011). It compliments but is BENN D.I., KIRKBRIDE M.P., OWEN L.A. & BRAZIER V. (2003) - Glaciated valley landsystems. In: Evans D.J.A. (Ed.), «Glacial Landsystems». clearly distinct from «paraglacial», the direction of influ- Arnold, London, 370-406. ence being reversed (Slaymaker, 2011). GLMs involve styles BOULTON G.S. & EYLES N. (1979) - Sedimentation by valley glaciers: A and combinations of ice-margin deposits and large-scale as- model and genetic classification. In: Schluchter C. (Ed.), «Moraines semblages. Not only are deposits of other processes interca- and Varves». Balkema, Rotterdam, 11-24. lated with glacial ones, but glacial action is directly influ- BURBANK D.W., LELAND J., FIELDING E., ANDERSON R.S., BROZOVIC N., enced by non-glacial processes. As such the GLMs com- REID M.R. & DUNCAN C. (1996) - Bedrock incision, rock uplift and prise a set of «transglacial landsystems». What distin- threshold hill slopes in the northwestern Himalayas. Nature, 379, guished them has been the influence of post-glacial devel- 505-510. opments along the upper Indus streams, notably the wide- CONWAY W.M. (1894) - Climbing and exploration in the Karakoram- spread occurrence of massive rock slope failures. The main Himalaya. Appleton, New York, 3 volumes, 315 pp. consequence has been chronic disturbance and fragmenting DELINE P. (2009) - Interactions between rock avalanches and glaciers in the Mont-Blanc massif during the late Holocene. Quaternary Science of the rivers leading, not only to intermontane fluvial and la- Reviews, 28, 11-12, 1070-1083. custrine sedimentation, but episodes of aggradation around DERBYSHIRE E. & OWEN L. (1990) - Quaternary alluvial fans in the Ka- glacier termini that reach the stream valleys (Hewitt, 2006). rakoram Mountains. In: Rachochi A.H. & Church M. (Eds.), «Allu- The GLMs are linked to episodes of intermontane sed- vial Fans: A Field Approach». Wiley, New York, 27-54. imentation in landslide-disturbed river reaches; not only, DERBYSHIRE E., JIJUN L., PERROT F.A., XU S. & WATERS R.S. (1984) - and not necessarily at all, to glacier fluctuations. There Quaternary glacial history of the Hunza Valley, Karakoram Mountains, seems to be, in part at least, a paraglacial influence of Pakistan. In: Miller K.J. (Ed.), «International Karakoram Project, v. 2». Cambridge University Press, Cambridge, 456-495. glacially over-deepened and steepened valley walls (He- EYLES N. (1983) - Glacial geology: An introduction for engineers and earth witt, 2009b). This supports an interpretation in which scientists. Pergamon, Oxford, 409 pp. widespread and recurrent interruption of pro-glacial rivers FORT M. (1987) - Sporadic morphogenesis in a continental subduction set- by post-glacial landslides, and related geomorphic re- ting: An example from the Annapurna range, Nepal Himalaya. Zeit- sponses, has been critical for the timing, intensity and scale schrift fur Geomorphologie (Supplementband), 63, 9-36. of GLM-building. Since landslides onto the ice and FORT M. (1995) - The Himalayan Glaciation: myth and reality. Journal processes affecting valley side troughs are important for Nepal Geological Society, 11, 257-272. the GLM assemblages, these tend to be transglacial FORT M. (2000) - Glaciers and mass wasting processes: Their influence on landsystems in a broader sense (Iturrizaga, 2011). the shaping of Kali Gandaki valley, Nepal. Quaternary International, 65-66, 101-119. FORT M. & PEULVAST J.-P. (1995) - Catastrophic mass movements and morphogenesis in the Peri-Tibetan ranges. Examples form west Kun CONCLUSIONS Lun, East Pamir and Ladakh. In: Slaymaker O. (Ed.), «Steepland Geomorphology». Wiley, Chichester, 171-198. Past work has interpreted GLM and related deposits as GARDNER J.S. & HEWITT K. (1989) - Surge of the Bualtar Glacier, Karako- driven by climate change and recording neoglaciation. ram Ranges, Pakistan: A possible landslide trigger. Journal of Glacio- There is no question that latero-terminal moraines and logy, 32,129-140. sediments assemblages found in the Karakoram do record GIARDINO J.R., REGMI N.R. & VITEK J.D. (2011) - Rock glaciers. In: Singh purely glacial fluctuations due to climate, mass balance V.P., Singh P. & Haritashaya U.K. (Eds.), «Encyclopaedia of Snow, change or surge activity. Countless examples have been Ice and Glaciers». Springer, Dordrecht, 943-948. observed during LIA fluctuations and subsequent ad- GOUDIE A.S. (1984) - The geomorphology of the Hunza Valley, Karakoram Mountains, Pakistan. In: Miller K.J. (Ed.), «International Karakoram vances, and in surge events. However, neither the direct Project», vol 1. Royal Geographical Society, London, 359-410 pp. nor indirect roles of glacier fluctuations explain the scale, HAEBERLI W. (1985) - Creep of mountain permafrost: Internal structure timing, complexities and ice levels peculiar to the major and flow of Alpine rock glaciers. Mitteilungen Versuchsanstalt fur GLMs. What is distinctive about these developments in- Vasserbau. Hydrol Glaziologie, 77, 1-142. volves post-glacial adjustments in Karakoram and sur- HASERODT K. (1984) - Aspects of actual climatic conditions and historic rounding mountain ranges. They are mainly driven by fluctuations of glaciers in western Karakorum. Journal Central Asia, slope instability, massive collapses, and interactions of 7-2, 77-93. large landslides with axial drainage systems. It seems likely HASERODT K. (1989) - Zur pleistozanen und postglazialen Vergletsche- that this, and GLMs, are typical of interglacial conditions, rungzwishen Hindukusch, Karakorum und Westhimalaya. In: Haserodt K. (Ed.), «Beitragen und Materialen zur Regionalen Geographie», but they should not be confused with more strictly glacial Volume 2. Institut für Geographie der Technischen Universitat, Ber- landsystems. lin, 182-233.

92 HEWITT K. (1969) - Glacier surges in the Karakoram Himalaya (Central KUHLE M. (2006) - The past Hunza glacier in connection with a Pleistoce- Asia). Canadian Journal Earth Sciences, 6-4, Part 2, 1009-1018. ne Karakoram Ice Stream Network during the Last Ice Age (Würm). HEWITT K. (1988) - Catastrophic landslide deposits in the Karakoram In: Kreutzmann H. (Ed.), «Karakoram in Transition: Culture, Deve- Himalaya. Science, 242, 64-67. lopment and Ecology in the Hunza Valley». Oxford University Press, Karachi, 24-48. HEWITT K. (1993) - Altitudinal organization of Karakoram geomorphic processes and depositional environments. In: Shroder J.F. Jr. (Ed), LI J., DERBYSHIRE E. & SHUYING X. (1984) - Glacial and paraglacial sedi- «Himalaya to the Sea: Geology, Geomorphology and the Quaternary». ments of the Hunza Valley, North-West Karakoram, Pakistan: A preli- Routledge, New York, 159-183. minary analysis. In: Miller K.J. (Ed.), «The International Karakoram Project», vol. 2. Cambridge University Press, Cambridge, 496-535. HEWITT K. (1998) - Catastrophic landslides and their effects on the Upper Indus streams, Karakoram Himalaya, northern Pakistan. Geomorpho- MACDONALD K.I. (1989) - Impacts of glacier-related landslides at Hopar, logy, 26, 47-80. Karakoram Himalaya. Annals of Glaciology, 13, 185-188. HEWITT K. (1999) - Quaternary moraines vs. catastrophic rock avalanches MASON K. (1930) - The Glaciers of the Karakoram and Neighbourhood. in the Karakoram Himalaya, Northern Pakistan. Quaternary Resear- Records of the Geological Survey of India, LXIII/2, Government of ch, 51-3, 220-237. India, Central Publications Branch, Calcutta, 214-278. HEWITT K. (2001) - Catastrophic rockslides and the geomorphology of the MEINERS S. (1998) - Preliminary results concerning historic to Post-glacial Hunza and Gilgit Basins, Karakoram Himalaya. Erdkunde, 55, 72-94. stages in the NW-Karakorum (Hispar Muztagh, Batura Muztagh, Hewitt K. (2002) - Postglacial landform and sediment associations in a Rakaposhi Range). In: Stellrecht I. (Ed.), «Karakoram-Hindukush- landslide-fragmented river system: the transHimalayan Indus streams, Himalaya: Dynamics of Change», 4/1. Rüdgers Köppe Verlag, Köln, northern Pakistan. In: Hewitt K., Byrne M.-L., English M. & Young 49-70. G. (Eds.), «Landscapes of Transition: Landform and Sediment Asso- MILLER K.J. (Ed.) (1984) - The International Karakoram Project. Cam- ciations in Cold Regions». Kluwer, Amsterdam, 63-91. bridge University Press, Cambridge, 2 volumes, 635 pp. HEWITT K. (2006) - Disturbance regime landscapes: mountain drainage sy- OUIMET W.B., WHIPPLE K.X., CROSBY B.T., JOHNSON J.P.H.J. & SCHILD- stems interrupted by large rockslides. Progress in Physical Geography, GEN T.F. (2008) - Epigenetic gorges in fluvial landscapes. Earth Surfa- 30-3, 365-393. ce Processes and Landforms, 33, 1993-2009. HEWITT K. (2009a) - Rock avalanches that travel onto glaciers: Disturbance OWEN L. (1994) - Glacial and non-glacial diamictons in the Karakoram regime landscapes, Karakoram Himalaya, Inner Asia. Geomorphology, Mountains and Westren Himalaya. In: Warren W.P. & Croots D. 103, 66-79. (Eds.), «The Formation and Deformation of Glacial Deposits». HEWITT K. (2009b) - Paraglacial rock slope failures, disturbance regimes Balkema, Rotterdam, 9-24. and transitional landscapes, Upper Indus Basin, northern Pakistan. In: OWEN L. (2006) - Quaternary glaciation. In: Kreutzmann H. (Ed.), «Ka- Knight J. & Harrison S. (Eds.), «Periglacial and Paraglacial Proces- rakoram in Transition: Culture, Development and Ecology in the ses and Environments». The Geological Society, London, Special Hunza Valley». Oxford University Press, Karachi, 12-23. Publications, 320, 235-255. OWEN L.A. & ENGLAND J. (1998) - Observations on rock glaciers in the HEWITT K., CLAGUE J.J. & DELINE P. (2011a) - Catastrophic rock slope Himalayas and Karakoram Mountains of northern Pakistan and India. failures and mountain glaciers. Encyclopaedia of Snow and Ice. Geomorphology, 26, 1-2, 199-214. Springer Verlag, Vienna, 112-126. OWEN L., FINKEL R.C.M. & CAFFEE M.W. (2002) - Timing of multiple HEWITT K., CLAGUE J.J. & GOSSE J. (2011b) - Rock avalanches and the glaciations during the late Quaternary in the Hunza valley, Karakoram pace of late Quaternary development of river valleys in the Karakoram Mountains, Northern Pakistan, defined by cosmogenic radionucleide Himalaya. Geological Society of America Bulletin, 123, 1836-1850. dating of moraines. Bulletin of Geological Society of America, 114-5, HUMLUM O., CHRISTINASEN H.H. & JULIUSSEN H. (2007) - Avalanche- 147-157. derived rock glaciers in Svalbard. Permafr. Periglac. Process., 18, 75-88. PARK S.K., SEEBER L., BISHOP M. & SHRODER J. (2001) - Erosion, Hima- ITURRIZAGA L. (2006) - Transglacial landforms in the Karakoram (Paki- layan geodynamics, and the geomorphology of metamorphism. GSA stan): A case study from Shimshal Valley. In: KREUTZMANN H. (Ed.), Today, 11, 4-9. «Karakoram in Transition: Culture, Development and Ecology in the PORTER S.C. (1970) - Quaternary glacial record in Swat, Hohistan and Hunza Valley». Oxford University Press, Karachi, 419 pp. west Pakistan. Bulletin of Geological Society of America, 81, 1421- ITURRIZAGA I. (2011) - Lateroglacial landform systems. In: Singh V.P., 1446. Singh P. & Haritashaya U.K. (Eds.), «Encyclopaedia of Snow, Ice SCHNEIDER H.J. (1959) - Zur diluvialen Geschichte des NW-Karakorum. and Glaciers». Springer, Dordrecht, 704-708. Mitteilungen Der Geographische Gesellschafte Munchen, 44, 201- JISKOOT H. (2011) - Glacier surging. In: Singh V.P., Singh P. & Hari- 216. tashaya U.K. (Eds.), «Encyclopaedia of Snow, Ice and Glaciers». SCHNEIDER H.J. (1969) - Minapin: Gletscher und Menschen im NW- Springer, Dordrecht, 415-428. Karakorum. Die Erde, 100, 266-286. KALVODA J. (1992) - Geomorphological record of the Quaternary Orogeny SEARLE M.P. (1991) - Geology and tectonics of the Karakoram mountains. in the Himalaya and the Karakoram. Developments in Earth Surface Wiley, New York, 358 pp. Processes, 3. Elsevier, Amsterdam, 315 pp. SHARP M.J. (1988) - Surging glaciers: Behaviour and mechanisms. Progress KALVODA J. & GOUDIE A.S. (2007) - Landform evolution in the Nagar re- in Physical Geography, 12-3, 349-370. gion, Hispar Mustagh Karakoram. In: Kalvoda J. & Goudie A.S. (Eds.), «Geomorphological Variations». Nakladatelstvi, Prague, 87-126. SHRODER J.F., OWEN L.A. & DERBYSHIRE E. (1993) - Quaternary Glacia- tion of the Karakoram and Nanga Parbat Himalaya. In: Shroder J.F. KICK W. (1989) - The decline of the last Little Ice Age in High Asia com- pared with that in the Alps. In: Oerlemans J. (Ed.), «Glacier Fluctua- (Ed.), «Himalaya to the Sea: Geology, Geomorphology and the Qua- tions and Climate Change». Kluwer Academic Publishers, Dordre- ternary». Routledge, London, 132-158. cht, 129-142. SHRODER J.F. JR., BISHOP M.P., SLOAN V. & COPLAND L. (2000) - Debris- covered glaciers and rock glaciers in the Nanga Parbat Himalaya, Paki- KUHLE M. (1990) - Ice marginal ramps and alluvial fans in semiarid moun- tains: convergence and divergence. In: Rachochi A.H. & Church M. stan. Geografiska Annaler 82A, 17-31. (Eds.) «Alluvial Fans: A Field Approach». Wiley, New York, 55-68. SLAYMAKER O. (2011) - Criteria to distinguish between periglacial, proglacial KUHLE M. (2004) - The Pleistocene Glaciation in the Karakoram Moun- and paraglacial environments. Quaestiones Geographicae, 30, 85-94. tains: reconstruction of past glacier extensions and ice thicknesses. VISSER P.C. (1928) - Von dem Gletschern am Oberstn Indus. Zeitschrift Journal Mt. Science. 1-3, 17-298. fur Glets., 16, 169-229.

93 VISSER C. & VISSER-HOOFT J. (1935-1938) - Karakorum wissenschaftliche WICHE E. (1961) - Klimamorphologische Unterschungen im westlichen ergebnisse der Niederlandischen expedition in den Karakoram und die Karakoram. Tagungsbericht u.wiss. Abh.Dt. Geographentag, Berlin Angrezenden Gebiete in den jahren 1922, 1925, 1929-1939, und 1935. 1959. Wiesbaden, 190-203. E.J. Brill, Leiden, 216 pp. WORKMAN F.B. (1908) - Further explorations in the Hunza-Nagar and the WHALLEY W.B. & MARTIN H.E. (1992) - Rock glaciers: II models and me- Hispar Glacier. Geographical Journal, 32, 495-496. chanisms. Progress in Physical Geography, 16-2, 127-186. ZEITLER P.K., MELTZER A.S., KOONS P.O., CRAW D., HALLET B., CHAM- VON WISSMANN H. (1959) - Die heutige Vergletscherung und Schneegrenze BERLAIN C.P., KIDD W.S.F., PARK S., SEEBER L., BISHOP M. & SHRO- in Hochasien mit Hinweisen die Vergletscherung der letzten Eiszeit. DER J. (2001) - Erosion, Himalayan geodynamics, and the geology of Akademie der Wissenschaften und der Literatur, Mainz, 1103-1431. metamorphism. GSA Today, 11, 4-8. WICHE E. (1958) - Die Oesterreichische Karakorum-Expeition 1958. Mit- teilungen Geographische Gesellschaft Wien, 100, 280-294. (Ms. received 30 June 2012; accepted 1 March 2013)

94